|
 |
|
|||||||
|
|||||||||
© 2003 American Public Health Association
LETTER |
Stephen W. Marshall and Charles Poole are with the Department of Epidemiology, School of Public Health, and Anna E. Waller is with the Department of Emergency Medicine, School of Medicine, University of North Carolina, Chapel Hill.
Correspondence: Requests for reprints should be sent to Stephen W. Marshall, PhD, CB 7435, McGarvan-Greenberg Hall, University of North Carolina, Chapel Hill, NC 27599-7435 (e-mail: smarshall{at}unc.edu).
In a recent meta-analysis on sunscreen use and malignant melanoma risk,1 Huncharek and Kupelnick noted the difficulties inherent in using the case–control design to study this topic. We agree with the authors that the previously observed positive associations (increased malignant melanoma risk with increased sunscreen use) may be largely due to various biases, and that these biases may be more pronounced in studies using hospital-based cases and controls.
We are concerned, however, that the meta-analytic summary odds ratio of 1.01 (95% confidence interval = 0.46, 2.28) computed from the subgroup of 4 studies that used general population controls may still be biased, potentially masking any protective effect of sunscreen use. The hypothesis that sunscreen use does not prevent malignant melanoma among persons whose skin is highly exposed to sunlight is implausible on the basis of animal and human models.
The fact that these 4 studies are not heterogeneous (on the basis of the Q statistic) does not mean that they are unconfounded. Potential biases include more vigilant detection of melanoma among sunscreen users, differential misclassification of sunscreen use between cases and controls, and confounding by socioeconomic status. These biases may still be present even in general-population case–control studies.
It is also important to note that none of these 4 studies comprehensively controlled for exposure to sunlight. Two of the 4 studies controlled for sunburn history, sunbathing history, or both, but none of them controlled directly for length of time exposed to sunlight. Exposure to sunlight is an important risk factor for malignant melanoma that is likely to be highly associated with sunscreen use, thereby creating the potential for enough upward confounding to obscure a protective sunscreen effect. The only study that attempted to control for number of hours of sun exposure, according to the meta-analysts’ Table 1, was a study with cases and controls from a single hospital that reported a relatively imprecise odds ratio estimate of 0.20 for sunscreen use (its 95% confidence interval was reported in Table 1 as 0.01, 0.80 and in Table 2 as 0.04, 0.79).
Finally, the authors failed to note that highly effective sunscreen preparations (sun protection factor [SPF] 20 and above) have become widely available on the international retail market only within the past 15 to 20 years. In fact, it wasn’t until 1979 that the Food and Drug Administration developed the SPF rating system. Since 2 of the 4 studies date from the 1980s, it is probable that the many of the subjects included in the meta-analysis had limited access to effective sunscreen preparations for much of their exposure history, particularly for exposures during childhood and adolescence.
We concur with the authors that, in the absence of conclusive evidence, use of sunscreens is a prudent cancer prevention strategy.
Reference
1. Huncharek M, Kupelnick B. Use of topical sunscreens and the risk of malignant melanoma: a meta-analysis of 9067 patients from 11 case–control studies. Am J Public Health.2002;92:1173–1177
This article has been cited by other articles:
|
|
 |
|
  A. Dupuy, A. Dunant, J.-J. Grob, and for Reseau d'Epidemiologie en Dermatologie Randomized Controlled Trial Testing the Impact of High-Protection Sunscreens on Sun-Exposure Behavior Arch Dermatol, August 1, 2005; 141(8): 950 - 956. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |